Inspection and classification of components assembly on PCBs applying Deep Learning techniques

Abstract

The Printed Circuit Board (PCB) is used in almost every electronic product we use everyday, whether for commercial purposes or in other technological applications. Due to the relevance of the application, the PCBs, after the component assembly process, need an inspection system and assembly defects location to guarantee the quality of their applications. Mistakenly mounting a board component can cause significant failures in the final product step. To classify the defects of the artificially generated components of the reference PCBs, the algorithm based on convolutional neural networks (CNNs) was applied. And the results indicated that the applied algorithm can be used in the inspection and classification of defects in PCIs for a low-cost system.